The purpose of this study, proposed by a new investigator, is to determine the multi-level factors associated with late antiretroviral therapy (ART) initiation (i.e., in the advanced stages of HIV disease) in areas of sub-Saharan Africa where ART is rapidly being scaled-up. Most patients in sub-Saharan Africa initiate ART late, resulting in substantial early mortality, more complicated and costly clinical management, and missed opportunities for prevention of HIV transmission. Without a better understanding of the determinants of late ART initiation, including the role of upstream pre-cursors (e.g., late diagnosis and late enrollment into care), efforts aimed at achieving more timely ART initiation will be severely hampered, and the full potential of ART scale-up in the region will not be achieved. The proposed study will consider factors relevant to three critical time points: diagnosis, enrollment into care, and ART initiation, and thus will be able to quantify the relative contribution of late diagnosis and late enrollment to HIV care to late ART initiation. Our ultimate goal is to inform the development of interventions to increase the likelihood of timely ART initiation. We propose a study that utilizes both quantitative and qualitative methodologies to examine the multi-level determinants of late ART initiation (defined as CD4<100 cells/5L or WHO stage IV) among patients initiating ART. These methodologies include retrospective analysis of existing service-delivery data, in-depth clinic and program assessments, and a case-control study. In Aim 1, we will examine the influence of factors at multiple levels-contextual (e.g., urban/rural, testing coverage, stigma), clinic (e.g., appointment adherence support, active testing, staffing), and individual (e.g., age, sex, entry point). We will draw on an existing dataset of routinely-collected clinic-level and individual-level service-delivery data on 72,007 persons who initiated ART from 71 HIV care and treatment clinics in 23 sub-regions of nine African countries. These clinic and individual- level data will be combined with contextual-level data from nationally representative household surveys to examine the multi-level determinants of late ART initiation. Prior studies have not examined these multiple levels of influence over such a wide range of clinics and settings nor have they examined longitudinal data from the pre-ART phase of care leading up to ART initiation. Aims 2 and 3 we be carried at four clinics with historically low, medium, and high rates of late ART initiation. We will conduct in-depth clinic and program assessments to ascertain clinic-level enablers/barriers to earlier ART initiation (Aim 2). Lastly, as very little is known about individual-level risk factors for late ART initiation, we will conduct detailed interviews as part of a case-control study examining the role of normative and behavioral factors (e.g., health beliefs, depression, social-support, substance use) through interviews of 360 cases who started ART late and 360 controls who started ART earlier (Aim 3). The overarching goal of this study is to provide critical information to inform strategies for achieving more timely ART initiation, ultimately reducing HIV-related morbidity and mortality.